in the loop

GPT-5.4 Pro
#13 of 274 scored.

OpenAIMar 5, 2026Closed

Capabilities Index

Range 156.6–161.8

158.9

Context

Takes text, image, file

1.05M

Input

Per 1M tokens

$30

Output

Per 1M tokens

$180

Benchmarks

How it scores.

  • GPQA Diamond92.8%

    6th of 186

  • FrontierMath82.5%

    13th of 81

  • ARC-AGI-283.3%

    13th of 79

  • Humanity's Last Exam41.5%

    5th of 40

  • SimpleQA Verified46.3%

    35th of 77

Everything else Epoch has for it

  • ARC-AGI94.5%
  • SimpleBench68.9%
  • FrontierMath-Tier-4-v2-Private58.5%
  • WeirdML57.4%
  • Chess Puzzles56.4%

On the index

Around it.

  1. #10GPT-5.6 SolOpenAI · $2.00 in161.7
  2. #11GPT-5.6 TerraOpenAI · $2.00 in159.6
  3. #12GPT-5.5OpenAI · $5.00 in159.1
  4. #13GPT-5.4 ProOpenAI · $30 in158.9
  5. #14Claude Opus 4.8Anthropic · $5.00 in158.2
  6. #15Kimi K3Moonshot · $0.80 in157.4
  7. #16Gemini 3.7 FlashGoogle · $0.75 in157.3

Sources